\t随着高性能计算的飞速发展与模型机理研究的不断深入，计算机仿真模型呈现结构复杂、高维、高度非线性的特点，同时仿真模型的运行时间也大幅增加，对复杂仿真模型的建模与分析提出了新的挑战。本书针对复杂模型计算耗时长、高维非线性等问题，以代理模型为技术途径进行实验设计、灵敏度分析、优化设计等方面的研究，并以船舶操纵性为建模对象进行应用。



\t本书可作为研究复杂系统建模与仿真的参考用书。

Chapter 1 Introduction
1.1 Surrogate Model
1.2 Design of Experiments
1.3 Global Sensitivity Analysis
1.4 Book Overview
Chapter 2 Optimal Latin Hypercube Design Using Local Search-based Genetic Algorithm
2.1 Optimal Latin Hypercube Design
2.2 Local Search-based Genetic Algorithm for LHD Optimization
2.3 Performance Comparison of Optimization Methods
2.4 Summary
Chapter 3 Active Learning of Multi-kernel Kriging Surrogate Models Using Regional Discrepancy and Space-filling Criteria
3.1 Formulation of Ensemble Surrogate Model
3.2 Ensemble Learning for Kriging Surrogate Models
3.3 Experimental Study
3.4 Summary
……